Glucokinase

In mammals glucose can be utilized by all cells to produce ATP and additionally in hepatocytes, myocytes and cardiomyocytes to form glycogen which serves as an energy store. Glucose enters into the cells through glucose transporters (GLUT) from which at least 5 have been identified. The first step of glucose utilization is phosphorylation of glucose to glucose-6-phosphate. This can be done by four different hexokinases I-IV from which hexokinases I-III have relatively high affinities for different hexose substrates and their molecular mass is about 100 kDa. The hexokinases I-III are inhibited by physiological concentrations of their reaction product glucose-6-phosphate (G6-P). In contrast, hexokinase IV (EC 2.7.1.2), also known as glucokinase (GK) with a mass of about 52 kDa exhibits and displays sigmoidal kinetics with a Hill coefficient of ~ 1.5 – 1.7 and is not inhibited by its reaction product glucose 6-phosphate. Even though glucose is the preferred substrate of the enzyme, GK can also convert mannose to 2- deoxyglucose and fructose (Lenzen et al., 1987). However, GK is inhibited by glucosamine and its derivatives, by mannoheptulose, alloxan, and in vitro by palmitoyl-CoA and other long chain acyl-CoA esters.

1.1.1 Glucose homeostasis

GK was first discovered in the liver (Grossbard and Schimke, 1966) and a few years later GK activity was also assigned to pancreatic cells (Matschinsky et al., 1968). Later it was shown that GK expression is not only restricted to hepatocytes and pancreatic β-cells, but is also present in some neuroendocrine cells of the gastrointestinal tract and the brain (Jetton et al., 1994). The hepatic GK exerts a very strong influence on glucose homeostasis by glucose utilization and glycogen synthesis. Small variations of GK levels in transgenic mice modulated blood glucose concentration (Hariharan et al., 1997). In addition, overexpression of GK in primary hepatocytes led to elevations in glucose 6-phosphate (O'Doherty et al., 1996), which as a central metabolite triggers increased rates of glycolysis and glycogen synthesis (Aiston et al., 1999). The importance of GK was underlined by the findings that homozygous knockout mice died a few days after birth and heterozygous knockout mice appeared normal at birth but died within 4 days due to a defect in insulin secretion in response to glucose and hyperglycaemia (Bali et al., 1995). In addition, β-cell-specific knockout mice were similar to global GK knockout mice showing hepatic steatosis, a depleted hepatic glycogen content and an ~70% decrease in plasma insulin concentrations suggesting the role of GK as glucose sensor in β-cells. In contrast, liver specific loss of hepatic GK was not lethal and had relatively small effects on

plasma glucose concentration. These results indicate a cooperative mode of both β-cell GK and hepatocyte GK in the regulation of glucose homeostasis.

In pancreatic β-cells, GK was considered to be the ‘glucose sensor’ (Garfinkel et al., 1984), as the phosphorylation of glucose within β-cells is tightly coupled to insulin secretion. When glucose enters the β-cell by type 2 glucose transporters it is phosphorylated by GK and converted to glucose-6-phosphate, which effectively traps glucose inside the cell. As glucose metabolism proceeds, ATP is produced in the mitochondria. This increase in the ATP:ADP ratio shuts ATP-gated potassium channels in the β-cell membrane, thus keeping positively charged potassium ions inside the β-cell. This decrease in K⁺-efflux depolarizes the β-cell, resulting in opening of voltage-gated calcium-channels, which in turn flood Ca²⁺ ions into the β-cell. The increase in Ca²⁺ ion concentration then triggers the secretion of insulin via exocytosis (Gilon et al., 2002).

Thus, even small changes in GK activity can be physiologically significant, as they directly affect the threshold for glucose-stimulated insulin secretion (Fig 1).

Figure 1. Role of glucokinase in glucose homeostasis and the interplay between pancreatic β-cells and hepatocytes In pancreatic β-cells, glucose is transported into the cells via glucose transporter Glut2, and phosphorylated by GK to yield glucose-6-phosphate (G6-P). Subsequently, glycolysis and mitochondrial metabolism increase the ATP:ADP ratio, thus leading to inactivation of the Kir6.2 potassium channel, and following a depolarisation of the membrane, to an influx of Ca²⁺ which triggers insulin secretion.

Glucose per se or a glucose metabolite, such as G6-P, can activate GK expression either transcriptionally, translationally or post-translationally. Insulin can also activate GK expression via a transcriptional mechanism.

In hepatocytes, insulin acts as the primary activator of GK transcription, the prerequisite for production of G6-P by GK and the storage of glucose as glycogen, which is stimulated by insulin. Conversion of glucose to G6-P with subsequent glycolysis also leads to acetyl-CoA formation which can be used for lipid synthesis.

1.1.2 Glucokinase regulatory protein (GKRP)

In the liver, GK activity and subcellular localization is regulated by a 68 kDa GK regulatory protein (GKRP) in conjugation with fructose 6-phosphate (F6-P) (Vandercammen and Van, 1990). GKRP inhibits GK, with respect to glucose, by forming a protein-protein complex which is sequestered in the nucleus. This effect can be reversed by high glucose or fructose 1-phosphate (F1-P) (Veiga-da-Cunha and van Schaftingen, 2002). Surprisingly, GK lacks a nuclear localization sequence and entry into the nucleus depends on GKRP by a piggy-back mechanism. By contrast, GK has a nuclear export signal (NES) sequence (300 ELVRLVLLKLV 310) near to its carboxy terminus that is masked upon binding of GKRP, thus ensuring that the GK-GKRP complex remains in the nucleus (Shiota et al., 1999). When blood glucose levels rise after feeding, GKRP is released from GK and allows export from the nucleus to the cytoplasm via an active process. Within the cytoplasm GK is catalytically active and converts glucose to glucose 6-phosphate. However, as the blood glucose levels begin to fall, GKRP binds cytoplasmic GK and moves it back into the nucleus.

GKRP ‘knockout’ experiments showed that homozygous GKRP^-/- mice have a ~ 40% reduction in liver GK protein levels and enzymatic activity. These deficient mice show an impaired glucose tolerance to a bolus of injected glucose, which is due to the inability to recruit GK from a nuclear reserve. These results strongly suggest the physiological role of the regulatory protein which may aid to provide a functional reserve of GK that can be quickly released after a meal.

However, it is questioned whether GKRP as such is present in the pancreatic β-cell.

Alternatively, although splice variants of GKRP have been described, it appears that other GK-binding partners exist, including long chain fatty acyl-CoA, propionyl-CoA carboxylase β-subunit precursor, insulin-containing granules and nitric oxide synthase (NOS), dual specific phosphatase (DUSP12), the bifunctional enzyme 6-phosphofructo-2-kinase/fructose-2-6-bisphosphatase (PFK-2/FBPase-2) and β-cell matrix proteins (Baltrusch et al., 2001).

1.1.3 Glucokinase protein structure

The resolved GK crystal structure allowed an easier elucidation of its unique kinetic properties.

The crystal structure of GK unveiled a palm shaped structure with a small and large domain separated by an interdomain cleft. The connecting region, Asn204 and Asp205, in conjunction with Glu256 and Glu290 of the large domain and Thr168 along with Lys169 of the small domain, are involved in glucose binding. In addition, an allosteric site was identified at the interface between the two domains and is surrounded by connecting region-I, the large domain (β1strand and α5 helix) and the small domain (α13 helix) (Kamata et al., 2004).

The GK can be found in two confirmations along with an intermediate confirmation. Without glucose, GK exists in its thermodynamically favorable wide-open inactive confirmation. Once bound to glucose, GK undergoes a conformational change and switches from the inactive wide-open to an intermediate active wide-open confirmation and then to the very active closed form. Upon conversion of glucose, G6-P and ADP will be released and GK returns via the open confirmation to the wide-open form. As both forms slowly equilibrate and the conformational change from the closed to the open form is faster than from wide-open to open and vice versa, it appears that a large portion of GK exists in the open form for some time (Fig 2). These conformational changes indicate that GK operates in two cycles: a fast cycle and a slow cycle. If glucose binds to the intermediate open form, GK immediately enters into the catalytic cycle; if not (e.g., low glucose concentration) GK returns to the wide-open form and uses the slow cycle. This shift between catalytic cycles explains the mnemonic mechanism and the sigmoidal saturation curve for glucose of GK, as well as its ability to regulate blood glucose levels in vivo.

Figure 2. Model of the two glucokinase reaction cycles

The glucose-unbound glucokinase (GK) exists in a wide-open confirmation. When glucose binds to it, GK undergoes confirmational changes from the wide-open to the close active confirmation via the intermediate open but active conformation. After conversion of glucose, GK returns to the open form, thereby releasing glucose 6-phosphate (G6-P) and ADP. At this stage, the enzyme may aquire a new glucose molecule and may undergo an immediate new reaction (fast cycle) or, as under low glucose concentrations, return to the wide-open inactive conformation.

1.1.4 Glucokinase gene

Glucokinase is encoded by a single gene in humans, rats and mice. The human GK gene is located on chromosome 7p13, whereas the mouse gene is located on chromosome 11. The gene consists of 10 exons and has two widely separated and functionally distinct cell-type-specific promoters. Cloning and characterization of an 83 kb clone (P1-305) having both promoters and all coding sequences of the mouse GK revealed that the 11 exons of total gene span 49 kb, with exons 1β (upstream promoter) and 1L (downstream promoter) being separated by ~35 kb (Postic et al., 1995). These alternate promoters in conjunction with the use of different first exons lead to tissue-specific GK expression. The β-cell GK mRNA is a product of exon 1a and 2-10, whereas the liver expresses GK from exons 1b and 2-10 and 1b, 1c and 2-10 due to alternate splicing.

The upstream (β-cell promoter) GK promoter is expressed specifically in several different cell types including pancreatic β-cells, neural/neuroendocrine cells of the gastrointestinal tract and corticotropes of the pituitary (Liang et al., 1991). However, the downstream GK promoter is mainly responsible for hepatic GK gene transcription. The upstream β-cell specific promoter has 70% sequence similarity among human, rat and mouse species.

1.1.5 Glucokinase promoter regulation

DNaseI hypersensitive analysis and DNaseI footprint assays of the liver specific GK promoter revealed that, both in rats and mice, the hepatic GK promoter encompasses ~7 kb of 5’ –flanking sequence (Postic et al., 1995). DNaseI footprinting assays identified seven elements A – G (in 3’

to 5’ direction), specifically interacting with liver enriched factors (Iynedjian et al., 1996). Recent studies have identified two binding sites for hepatocyte nuclear factor-6 (HNF-6, the prototype of the ONECUT family of transcription factors), located at the most distal (-7613/-7622) and the more proximal (-877/-868) sites of the promoter (Lannoy et al., 2002). Further, analysis of the more proximal part of the liver specific promoter showed the existence of additional footprints named as P2 (-87/-80) and P1 (-54/-35) (Iynedjian, 1998a). The element P2 has been shown to be bound by the basic helix-loop-helix transcription factors, upstream stimulatory factor-1 and -2 (USF-1, -2) (Iynedjian, 1998b), as well as by hypoxia-inducible factor-1 (HIF-1) (Roth et al., 2004b). The P1 element was identified to contain a binding site for hepatocyte nuclear factor-4α (HNF-4 α) (Roth et al., 2002).

The liver tissue exhibits a metabolic zonation and a zonated gene expression. The zonated pattern may be the result of the gradients in nutrients, hormones and oxygen which are formed due to the blood flow through the sinusoids and the metabolism of the cells. Especially oxygen

or insulin mainly contributes to the zonated gene expression in liver. The liver GK gene expression is mainly stimulated by insulin in the hypoxic perivenous zone of the liver (Iynedjian et al., 1989), (Krones et al., 2000), involving HIF-1 and HNF-4 to play an important role during this process. Previous results have shown that the cooperation between HIF-1, HNF-4 and the co-activator p300 contributed to the insulin-dependent GK induction (Roth et al., 2004a). The insulin signalling pathway involved in the GK gene expression was the PI3K/protein kinase B pathway, which also regulates the insulin-dependent expression of several other genes required for carbohydrate metabolism such as glucose 6-phosphatase (G-6-Pase), glucose transporter-1 (GLUT-1) and sterol regulatory element binding protein-1 (SREBP-1) (Iynedjian PB et al., 2000).

Very recent reports implicated sterol regulatory element binding protein (SREBP) as an insulin-dependent activator for transcription of hepatic GK. It has been shown that GK promoter activity was induced by SREBP-1a, and found that SREBP-1c binds to two sterol response elements designated (SREa -205/-197) and (SREb -183/-174) (Kim et al., 2004). However, the mutual relationship between the SREa and SREb is not well understood in the SREBP-1c-mediated activation of liver GK by insulin.

Furthermore, FoxO transcription factors are important targets of insulin signaling and contribute to the regulation of cell growth, differentiation, and metabolism. Insulin has a dynamic effect on the localization of FoxO. PKB/Akt pathway of insulin signaling inactivates the transactivation and promotes nuclear exclusion of FoxO. FoxO proteins exert both positive and negative effects on gene expression. Studies with adenoviral vectors in isolated hepatocytes reveal that FoxO1 stimulates gluconeogenic genes (PEPCK) and suppress glycolytic genes including glucokinase and SREBP-1c (Zhang et al., 2006). However, FoxO1-mediated repression of glucokinase gene was not understood completely.